With this post I’m announcing availability
of a new Python parser based on
LR(1) and GLR algorithms.
For years I’ve been working with and
developing PEG based
parsers (see Arpeggio
and textX). PEGs have some really nice
properties: they represent a concise description of top-down parsers and are
easy to understand and debug. Furthermore, with choice operator being ordered
PEG grammars can’t be ambiguous. Backtracking is natural and easy to implement
giving it unlimited lookahead. A composition of PEG grammars will give a valid
PEG grammar, although
it
might not be what you would expect.
I’m a long time vim user and I just love vim modal editing approach. But, I
always admired Emacs extensibity, introspection and self-documenting
capabilities. Although I gave Emacs a try couple of times in the past I just
couldn’t get used to non-modal editing and cumbersome finger-stretching keyboard
bindings. For some time I was keeping an eye
on spacemacs, emacs distribution geared towards vim
users. It looked like a perfect blend of vim and emacs so I gave it a try.
Although spacemacs is still beta the switch-over was surprisingly smooth. It
took me just a few days to get to the speed thanks
to Evil mode, a vim-like modal editing
support which is enabled by default in spacemacs. Evil works remarkably good,
even some often used plugins are ported like, for
example, surround.
Well, you know what they say: “You aren’t making a DSL tool if you haven’t
implemented Martin Fowler’s State
Machine language using your tool yet” ;)
This is a well known DSL example from his book Domain-Specific
Languages. If you are doing stuff
in the field of DSLs you should definitely read this book. There is an excerpt
from the book
available on-line so I advise you to read it before watching the video below.
Finally got some time to record first video for
textX. In this video I will show you how to
install textX and do some first steps. This is my first video tutorial so
please bear with me ;)
Today, while hacking on pyFlies I
run into trouble with the code generated for the
PsychoPy backend. PsychoPy is using
Pyo to generate sound. The problem
is that Pyo do not support
PulseAudio, only
JACK. This might be a good choice since JACK is
better suited for low latency sound processing, but today most linux distro use
PulseAudio which is better suited for general desktops.
In order to use PsychoPy, first thing to do is to start JACK. Luckily,
there are nice GUI apps to do that easily, e.g.
QjackCtl or
cadence. Starting JACK
will suspend PulseAudio while JACK is runnning.
Now, we must instruct pyo called from PsychoPy to use JACK. Although,
there is audio parameter in Server call of pyo lib which could be used to
specify JACK as the backend, this would require changing source code of
PsychoPy. There is luckily a non-obtrusive way to achieve the same effect.
Just specify environment variable:
export PYO_SERVER_AUDIO=jack
Now pyo lib will use JACK by default and the PsychoPy sound will work.